Certain investigation on lung cancer segmentation and stage classification using genetic and ai algorithms in CT images

Abstract

In recent years, imaging devices have been widely used in many newlinemedical fields to improve imaging in early diagnosis and treatment, especially newlinein many cancers such as cancer and cancer where time is important to newlineimprove performance. Check for defects in the target image. Diagnostic newlineimaging analytics is becoming increasingly common in the medical newlineprofession, particularly in non-invasive therapy and clinical examination. newlineOnly in the early stages of cancer can it be effectively treated, yet diagnosing newlinecancer in the early stages is challenging. Machine learning methods, artificial newlineintelligence, and deep learning algorithms can be utilized to categorize newlinebenign, malignant, and normal lung nodules in this scenario. newlineA genetic algorithm was used to create a more accurate newlinesegmentation and classification model for lung cancer. The input CT images newlinein this model are first preprocessed using the adaptive median filter and newlineaverage filter. The filtered images are histogram equalized, and the ROI of newlinecancer tissues is segmented using the Guaranteed Convergence Particle newlineSwarm Optimization method. Probabilistic Neural Networks (PNN) - based newlineclassification is used to categories images. The LIDC-IDRI (Lung Image newlineDatabase Consortium-Image Database Resource Initiative) benchmark dataset newlineand CT lung images were used as input for modelling experiments in newlineMATLAB (Matrix Labs). The results demonstrate that the proposed model newlineoutperforms existing methods by delivering precise classification outcomes newlinequickly. newline newline